GitLab

* add model files etc for MobileNetV2

rename files for MobileNetV1

initial implementation of MobileNetV1

fix conversion script

cleanup

write docs

tweaks

fix conversion script

extract hidden states

fix test cases

make fixup

fixup it all

remove main from doc link

fixes

fix tests

fix up

use google org

fix weird assert

* fixup

* use google organization for checkpoints

* run slow test on GPU

* remove unnecessary device assignment

* use `torch_device` instead

* Add LayerScale to NAT/DiNAT.

Completely dropped the ball on LayerScale in the original PR (#20219).
This is just an optional argument in both models, and is only activated for larger variants in order to provide training stability.

* Add LayerScale to NAT/DiNAT.

Minor error fixed.
Co-authored-by: Ali Hassani <ahassanijr@gmail.com>

* Update _toctree and clone original content

* Translate first three sections

* Add more translated chapters. Only 3 more left.

* Finish translation

* Run style from doc-builder

* Address recommended changes from reviewer

* zh quicktour(#20095)

* add zh to doc workflow

* remove untranslation from toctree
Co-authored-by: BeifangSusu <BeifangSusu@bfss.com>

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix device issue
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Add DiNAT

* Adds DiNAT + tests

* Minor fixes

* Added HF model

* Add natten to dependencies.

* Cleanup

* Minor fixup

* Reformat

* Optional NATTEN import.

* Reformat & add doc to _toctree

* Reformat (finally)

* Dummy objects for DiNAT

* Add NAT + minor changes

Adds NAT as its own independent model + docs, tests
Adds NATTEN to ext deps to ensure ci picks it up.

* Remove natten from `all` and `dev-torch` deps, add manual pip install to ci tests

* Minor fixes.

* Fix READMEs.

* Requested changes to docs + minor fixes.

* Requested changes.

* Add NAT/DiNAT tests to layoutlm_job

* Correction to Dinat doc.

* Requested changes.

* future proof our tf code

* parse tf versions

* remove double brackets

* oops get other bracket

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* [Proposal] Breaking change `zero-shot-object-detection` for improved
consistency.

This is a proposal to modify the output of `zero-shot-object-detection`
to provide better alignment with other pipelines.

The output is now strictly the same as `object-detection` whereas before
it would output lists of lists.

The name `candidate_labels` is used throughout for consistency with
other `zero-shot` pipelines.

The pipeline is changed to `ChunkPipeline` to support batching cleanly.

This removes all the lists and list of lists shenanigans, it's now a
matter of the base pipeline handling all this not this specific one.

**Breaking change**: It did remove complex calls potentials `pipe(images = [image1, image2],
text_queries=[candidates1, candidates2])` to support only
`pipe([{"image": image1, "candidate_labels": candidates1}, {"image": image2, "candidate_labels": candidates2}])`
when dealing with lists and/or datasets.
We could keep them, but it will add a lot of complexity to the code
base, since the pipeline is rather young, I'd rather break to keep the
code simpler, but we can revert this.

**Breaking change**: The name of the argument is now `image` instead of
`images` since it expects by default only 1 image. This is revertable
like the previous one.

**Breaking change**: The types is now simplified and flattened:

`pipe(inputs) == [{**object1}, {**object2}]`
instead of the previous
`pipe(inputs) == [[{**object1}, {**object1}], [{**object2}]]`
Where the different instances would be grouped by candidate labels
within lists.
IMHO this is not really desirable, since it would output empty lists and
is only adding superflous indirection compared to
`zero-shot-object-detection`.

It is relatively change free in terms of how the results, it does change
computation however since now the batching is handled by the pipeline
itself. It **did** change the results for the small models so there
seems to be a real difference in how the models handle this.

* Fixing the doctests.

* Behind is_torch_available.

* Add AnyPrecisionAdamW optimizer

* Add optim_args argument to TrainingArgs

* Add tests for AnyPrecisionOptimizer

* Change AnyPrecisionAdam default params to float32

* Move default_anyprecision_kwargs in trainer test

* Rename AnyPrecisionAdamW

* Add padding transformation

* Add in upstream changes

* Update tests & docs

* Code formatting tuples in docstring

* [ASR Examples] Update README for seq2seq

* add language info

* add training results

* re-word

* fix the doc to specify that add_prefix_space = False

* add correct expected output

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

- simplifies the devce checking test

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Add ResNetBackbone

* Define channels and strides as property

* Remove file

* Add test for backbone

* Update BackboneOutput class

* Remove strides property

* Fix docstring

* Add backbones to SHOULD_HAVE_THEIR_OWN_PAGE

* Fix auto mapping name

* Add sanity check for out_features

* Set stage names based on depths

* Update to tuple
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>

* Add docstrings for canine model

* Update CanineForTokenClassification
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

set the default cache_enable to True, aligned with the default value in pytorch cpu/cuda amp autocast (#20289)
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

* Fixing the doctests failures.

* Fixup.

* move contrastive search test to slow

* test hub tf callback

* create repo before cloning it

* Image transforms functionality used instead

* Import torch

* Import rather than copy

* Update src/transformers/models/conditional_detr/feature_extraction_conditional_detr.py

* Adding doctest for `object-detection` pipeline.

* Removed nested_simplify.

* Adding `zero-shot-object-detection` pipeline doctest.

* Remove nested_simplify.

* add warning on 8-bit models

- added tests
- added wrapper

* move to a private attribute

- remove wrapper
- changed `save_pretrained` method

* Apply suggestions from code review
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix suggestions
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* update part of the doc

* add temp values, fix part of the doc

* add template outputs

* add correct models and outputss

* style

* fixup

* Update configuration_deformable_detr.py comment

* Add DeformableDetrConfig to documentation_tests.txt